Numerous bug fixes and improvements

Greatly improved stability and robustness
Expanded documentation
Greatly improved logging for easier debugging
Improved terminal output
Add 1-node instant cluster option to lpjs ad-hoc
Disambiguate UI
Save/restore node states across daemon restart
Improve munge key distribution
Add built-in support for file transfer for nodes with no NFS
Enforce real memory limits
Improve service scripts
Standardize format of log messages
Record basic job history, add lpjs history command to view
Require same version of dispatchd and compd for checkin
Add lpjs peak-mem command to show RSS for completed jobs
Add #lpjs path to set path prior to file transfer

Add man pages, fix systemd script

Added a first draft of all man pages and improved the PDF guide
Fixed an issue with lpjs_dispatchd pid file required by systemd

Mostly functional

Most of the programs and scripts are reasonably functional, though bugs do remain
Next task will be documentation

CLI for node state, code cleanup

Add "lpjs node pause|up all|node-list"
Fix several bugs
Improve code readability

Add cancel, numerous improvements

Add an "lpjs cancel" subcommand with fairly robust process hunting an termination
Numerous bug fixes and other improvements

Pre-release with somewhat reliable job scheduling

This is the first pre-release, with seemingly reliable job scheduling and node status monitoring. Minor increments will occur as we improve basic functionality and code quality. When the system is deemed ready for general use, we will issue the first full release, 0.1.0.